Conifer transpiration in a montane watershed : environmental controls and methodological uncertainties
Looker, Nathaniel Thomas.
MetadataShow full item record
Transpiration, the evaporative loss of water from vegetation through stomata, is intertwined with biological, meteorological, and hydrologic processes from the scale of leaves to continents. Among the methods of estimating transpiration, thermometric sap flow sensors have gained favor due to their high temporal resolution and capacity to integrate across sub-canopy variability. Heterogeneity in growing conditions through space and time may contribute substantial uncertainty to sap flow-based inferences; hence, quantification of spatiotemporal variability in sap flow dynamics and plant physiological traits is required to improve confidence in transpiration estimates. In this study, I employ sap flow sensors to address the environmental sensitivity of water use dynamics in conifers growing in contrasting topographic positions (elevations, aspects, and hillslope positions) within a watershed; additionally, I explore how variability in sapwood traits may contribute to uncertainty in sap flow-based estimates of whole-tree transpiration. To these ends, I measured sap flow in 32 trees of 5 species throughout the 2014 growing season and estimated sapwood traits in those individuals and in co-occurring trees. I related sap flow dynamics to environmental variables (vapor pressure deficit and soil moisture) through lag analyses and nonlinear regression. A Monte Carlo-based simulation and mixed effects variance decomposition served to quantify the sensitivity of transpiration to variability in sapwood traits and the magnitude of that variability across ecological scales. The response of sap flow to vapor pressure deficit varied with species and across topographic positions, and sap flow dynamics diverged spatially as soil moisture declined during the summer. The simulation suggested that a common assumption (constant wood thermal diffusivity) could impart up to 100% overestimation of transpiration in trees with sapwood properties similar to those I observed. Taken together, these findings underscore the importance of sampling sap flow and sapwood traits at the spatiotemporal scale for which inferences are to be drawn.